Simply creating an LLM judge prompt isn't enough. Before deploying it, you must test its alignment with human judgment. Run the judge on your manually labeled data and analyze the results in a confusion matrix. This helps you see where it disagrees with you (false positives/negatives) so you can refine the prompt and build trust.

For subjective tasks, refining instructions has diminishing returns. The most effective way to improve AI performance is to provide it with a set of high-quality examples of the desired output. A library of five great examples is more powerful than a perfectly crafted prompt.

When using LLMs to judge other models' output, they consistently rate towards the middle of the curve, akin to humans giving a generic "7 out of 10." These AI judges are not "spiky" enough, failing to recognize unique or exceptional qualities that a human evaluator with strong taste would identify.

Using one LLM to evaluate another's output ("LLM as a Judge") is a common but deceptively difficult technique. Chip Huyen highlights that companies can spend up to 80% of their development time just writing and refining the complex evaluation guidelines for the judge LLM.

Standard benchmarks are insufficient. A more effective evaluation method is a hybrid approach, weighting a human's qualitative 'taste test' (e.g., 70%) more heavily than an LLM judge's automated score (e.g., 30%). This prioritizes subjective qualities like design, usability, and writing style.

'Taste' is a collection of specific preferences, not an abstract feeling. Document what makes an output 'good' by creating universal rules (e.g., 'write at a ninth-grade level,' 'avoid cheesy quotes,' 'no em dashes'). Feeding these documented rules to an AI transforms your subjective taste into repeatable instructions for consistent results.

Do not blindly trust an LLM's evaluation scores. The biggest mistake is showing stakeholders metrics that don't match their perception of product quality. To build trust, first hand-label a sample of data with binary outcomes (good/bad), then compare the LLM judge's scores against these human labels to ensure agreement before deploying the eval.

The concept of "taste" is demystified as the crucial human act of defining boundaries for what is good or right. An LLM, having seen everything, lacks opinion. Without a human specifying these constraints, AI will only produce generic, undesirable output—or "AI slop." The creator's opinion is the essential ingredient.

To avoid generic AI-generated text, use the LLM as a critic rather than a writer. By providing a detailed style guide that you co-created with the AI, its feedback on your drafts becomes highly specific and aligned with your personal goals, audience, and tone.